On-satellite processing method and system for low-orbit meteorological satellite
Through the combination of data exchange, preprocessing, resampling and product generation modules, combined with neural networks and dynamic backup, the problem of multi-channel processing of meteorological satellites is solved, real-time data processing and product generation of low-orbit meteorological satellites is realized, and emergency response capabilities are improved.
Patent Information
- Application Number
- CN202510392630.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-31
- Publication Date
- 2025-08-15
AI Technical Summary
The prior art is difficult to adapt to the needs of multi-channel and multi-load parallel processing of meteorological satellites, and cannot achieve efficient on-site data processing and product generation.
The data exchange module, preprocessing module, resampling module and product generation module are adopted to combine data through a unified format framework to separate, depacket, verify, position, calibration and resample, combine with a fully connected neural network to generate secondary products, and realize module redundancy through dynamic backup modules, and use power modules to provide stable power support.
Real-time processing and product generation of low-orbit meteorological satellite data has been realized, which alleviates the problem of insufficient transmission bandwidth and timeliness and ground, and is suitable for emergency rescue, disaster prevention and mitigation and other fields.
Smart Images

Figure CN120491113A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of payload design technology, and in particular to an onboard processing method and system for a low-orbit meteorological satellite. Background Art
[0002] Meteorological satellites are the most widely used and most cost-effective of my country's civilian remote sensing satellites. Their primary application areas include weather forecasting, climate change research, disaster prevention and mitigation, ecological and environmental monitoring, water conservancy, forestry, grasslands, aviation, and transportation. With the advancement of meteorological satellite technology, these satellites are shifting from providing traditional observational data to directly providing products. Among the world's polar-orbiting satellites, my country's new generation of polar-orbiting meteorological satellites is planned to be the first to be equipped with onboard processing systems. These satellites can generate relevant products directly in orbit, significantly shortening emergency response times, improving response capabilities, reducing disaster losses, and enabling rapid and precise resource allocation, thus promoting the modernization and scientific development of meteorological satellite applications.
[0003] Patent document CN115240053A discloses a small, general-purpose onboard real-time intelligent processing device for payload data. The device comprises an interface and control module, each connected to the satellite platform's integrated electronics and payload unit; at least one high-performance computing module, connected to the interface and control module and configured to provide high-performance parallel computing services; the module includes an FPGA processing unit, a DSP chip, and an AI processor, capable of providing an appropriate architecture based on bus instructions; and a power management module, connected to the satellite's power supply system to provide secondary power.
[0004] However, although patent document CN115240053A has a certain degree of versatility, the number of interfaces and modules is relatively simple, and cannot adapt to the requirements of multi-channel and multi-payload parallel processing of meteorological satellites. Summary of the Invention
[0005] In view of the defects in the prior art, the purpose of the present invention is to provide an on-board processing method and system for low-orbit meteorological satellites.
[0006] According to the present invention, a low-orbit meteorological satellite onboard processing method includes:
[0007] Step S1: The data exchange module receives and caches satellite multi-payload data and satellite platform data;
[0008] Step S2: The data exchange module combines the cached payload data and platform data using a unified format framework;
[0009] Step S3: The pre-processing module receives the combined data, separates, unpacks and verifies the data, and locates and calibrates the observation points therein;
[0010] Step S4: The data exchange module receives the positioning and calibration results and forwards them to the resampling module;
[0011] Step S5: The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time according to the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module;
[0012] Step S6: The data exchange module receives the processing result of the resampling module and forwards it to the product generation module;
[0013] Step S7: the product generation module performs product inversion on the processing results to generate various secondary products, and forwards the results to the data exchange module;
[0014] Step S8: The satellite platform sends a remote control command to specify the type of product to be sent, and the data exchange module selects the corresponding product data for output.
[0015] Preferably, step S2 includes encapsulating the payload data into packets of 144 bytes, each packet containing a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data and a 4-byte terminator.
[0016] Preferably, step S3 includes:
[0017] Step S3.1: Detect the synchronization word of the received data stream and enter the synchronization state after the synchronization word is detected;
[0018] Step S3.2: Identify the payload according to the payload identifier after the synchronization word, and determine whether there is packet loss based on the packet count;
[0019] Step S3.3: Check the correctness of the payload data format. If the correctness is found, the payload data is classified and stored.
[0020] Preferably, the generation of a standard time-space geographic grid includes: using satellite bus data as a time reference, assuming that the time sequence in the bus data output by the satellite is T1, T2, ..., and the first time received is T k , then T k The time starting point of the time-space reference grid is used as the satellite position corresponding to the first time as the spatial location, and the spatial location is calculated every ΔT time interval after the first time; the time interval ΔT is the minimum value of the imaging period of each satellite payload;
[0021] The resampling process of the positioning and calibration results on the geographic grid includes: for each observation point, selecting 16 time-space reference points around the observation point for consideration, determining the point with the smallest distance, and associating the observation point with the point with the smallest distance; after the calculation is completed, checking the calibration values a1, a2, ..., a of the points associated with each time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
[0022]
[0023] Preferably, step S7 includes: using a fully connected neural network, the fully connected neural network includes an input layer, a hidden layer and an output layer, wherein the number of hidden layers is 20; the fully connected neural network is trained using historical observations to obtain weights at each level;
[0024] The step S8 includes: the data exchange module can receive 64 products, and uses a 64-bit register to store product status. If each bit value is 1, the current product is issued; if each bit value is 0, the current product is not issued.
[0025] According to the present invention, an onboard processing system for a low-orbit meteorological satellite comprises: a data exchange module, a preprocessing module, a resampling module and a product generation module;
[0026] The data exchange module receives and caches satellite multi-payload data and satellite platform data;
[0027] The data exchange module combines the cached payload data and platform data using a unified format framework;
[0028] The pre-processing module receives the combined data, separates, unpacks and verifies the data, and locates and calibrates the observation points.
[0029] The data exchange module receives the positioning and calibration results and forwards them to the resampling module;
[0030] The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time based on the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module;
[0031] The data exchange module receives the processing results of the resampling module and forwards them to the product generation module;
[0032] The product generation module performs product inversion on the processing results to generate various secondary products and forwards the results to the data exchange module;
[0033] The satellite platform sends remote control instructions to specify the type of product to be issued, and the data exchange module selects the corresponding product data for output.
[0034] Preferably, the unified format framework includes encapsulating the payload data into packets of 144 bytes, each packet including a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data, and a 4-byte terminator;
[0035] It also includes: a dynamic backup module and a power supply module. The dynamic backup module realizes redundancy function and replaces the functions of the preprocessing module, resampling module and product generation module by uploading different software, and is used in the event of failure of the corresponding module; the power supply module realizes the conversion of the satellite's primary power supply into secondary power supply and supplies power to each module.
[0036] Preferably, the separation, unpacking and verification processing includes: performing synchronization word detection on the received data stream, entering the synchronization state after detecting the synchronization word, then determining the payload according to the payload identifier after the synchronization word, and judging whether there is packet loss in combination with the packet count, and then checking the correctness of the payload data format. After the check is passed, the payload data is classified and stored.
[0037] Preferably, the generation of a standard time-space geographic grid includes: using satellite bus data as a time reference, assuming that the time sequence in the bus data output by the satellite is T1, T2, ..., and the first time received is T k , then T k The time starting point of the time-space reference grid is used as the satellite position corresponding to the first time as the spatial location, and the spatial location is calculated every ΔT time interval after the first time; the time interval ΔT is the minimum value of the imaging period of each satellite payload;
[0038] The resampling process of the positioning and calibration results on the geographic grid includes: for each observation point, selecting 16 time-space reference points around the observation point for consideration, determining the point with the smallest distance, and associating the observation point with the point with the smallest distance; after the calculation is completed, checking the calibration values a1, a2, ..., a of the points associated with each time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
[0039]
[0040] Preferably, the generation of various secondary products adopts a fully connected neural network, the fully connected neural network comprises an input layer, a hidden layer and an output layer, wherein the number of hidden layers is 20; the fully connected neural network is trained using historical observations to obtain weights at each level;
[0041] The data exchange module can receive 64 types of products and uses a 64-bit register to store product status. If each bit value is 1, the current product is issued; if each bit value is 0, the current product is not issued.
[0042] The data exchange module uses anti-fuse FPGA and large-capacity FPGA.
[0043] Compared with the prior art, the present invention has the following beneficial effects:
[0044] The present invention realizes the on-orbit real-time processing and product generation of low-orbit meteorological satellite data, effectively alleviating the current problems of insufficient bandwidth and timeliness of satellite-to-ground transmission, and can be used in emergency rescue, disaster prevention and mitigation and other fields that have high requirements for the timeliness of satellite applications. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:
[0046] Figure 1 It is a schematic flow chart of the working method of the present invention;
[0047] Figure 2 It is a block diagram of the system composition of the present invention. DETAILED DESCRIPTION
[0048] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.
[0049] Example 1
[0050] According to the present invention, a low-orbit meteorological satellite on-board processing method is provided. Figure 1 Shown, including:
[0051] Step S1: The data exchange module receives and caches satellite multi-payload data and satellite platform data.
[0052] Step S2: The data exchange module combines the cached payload data and platform data using a unified format framework. Step S2 includes encapsulating the payload data into 144-byte packets, each of which contains a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data, and a 4-byte terminator.
[0053] Step S3: The pre-processing module receives the combined data, separates, unpacks and verifies the data, and locates and calibrates the observation points. Step S3 includes:
[0054] Step S3.1: Perform synchronization word detection on the received data stream, and enter the synchronization state after the synchronization word is detected.
[0055] Step S3.2: Determine the payload according to the payload identifier after the synchronization word, and determine whether there is packet loss in combination with the packet count.
[0056] Step S3.3: Check the correctness of the payload data format. If the correctness is found, the payload data is classified and stored.
[0057] The positioning and calibration include determining the observation longitude, latitude and altitude (ie positioning) and corresponding physical quantities (ie calibration) of each observation point, and forwarding the positioning and calibration results to the data exchange module.
[0058] Step S4: The data exchange module receives the positioning and calibration results and forwards them to the resampling module.
[0059] Step S5: The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time based on the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module. The specific method for generating the standard spatiotemporal geographic grid is as follows: using satellite bus data as the time reference, assuming that the time series in the bus data output by the satellite is T1, T2, ..., and after the onboard processing system is turned on, the first time it receives is T k , then T k The time starting point of the space-time reference grid is used as the corresponding satellite position as the spatial location. The spatial position is then calculated every ΔT time interval, where ΔT is the minimum value of the imaging cycle of each satellite payload.
[0060] The specific method of resampling the positioning and calibration results on the geographic grid is as follows: for each observation point, select 16 time-space reference points around it for consideration, determine the point with the smallest distance, and associate the observation point with the point with the smallest distance. For each time-space reference point, store up to 10 observation points associated with it. After the calculation is completed, check the calibration values a1, a2, ..., a of the points associated with each time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
[0061]
[0062] Step S6: The data exchange module receives the processing result of the resampling module and forwards it to the product generation module.
[0063] Step S7: The product generation module performs product inversion on the processing results, generates various secondary products, and forwards the results to the data exchange module. The specific method for generating various secondary products is to use a fully connected neural network comprising an input layer, a hidden layer, and an output layer, with 20 hidden layers. The network is trained using historical observations to obtain weights at each level. During use, the network input is the resampled values of each load channel, and the output is the predetermined inversion product.
[0064] Step S8: The satellite platform sends a remote control command specifying the product type to be delivered, and the data exchange module selects the corresponding product data for output. The data exchange module selects the corresponding product data for output as follows: the data exchange module can receive 64 types of products and uses a 64-bit register to store product status. If each bit value is 1, the product is delivered; if each bit value is 0, the product is not delivered.
[0065] In the above steps, data transfer between all modules is implemented by the data exchange module. Therefore, the data exchange module is the core module of the entire system and is usually implemented using anti-fuse FPGA and high-capacity FPGA in design.
[0066] The present invention aims to provide a universal, reconfigurable, low-power, and highly reliable engineering solution for on-board processing requirements of meteorological satellites.
[0067] Example 2
[0068] The present invention also provides an on-board processing system for a low-orbit meteorological satellite. Those skilled in the art can implement the on-board processing system for a low-orbit meteorological satellite by executing the step flow of the on-board processing method for a low-orbit meteorological satellite, that is, the on-board processing method for a low-orbit meteorological satellite can be understood as a preferred implementation of the on-board processing system for a low-orbit meteorological satellite.
[0069] According to the present invention, a low-orbit meteorological satellite on-board processing system is provided. Figure 2 As shown, it includes: a data exchange module, a preprocessing module, a resampling module, a product generation module, a dynamic backup module, and a power supply module. The data exchange module is responsible for system data management, enabling on-demand exchange of external and internal data. The preprocessing module performs unpacking, verification, positioning, and calibration of meteorological payload data. The resampling module unifies the spatial dimensions of data from different payloads and channels. The product generation module performs inversion of various application products. The dynamic backup module provides redundancy, allowing different software to replace the functions of the preprocessing module, resampling module, and product generation module in the event of module failure. The power supply module converts the satellite's primary power supply into a secondary power supply and provides power to each module. The data exchange module is implemented using an antifuse FPGA and a high-capacity FPGA.
[0070] According to the present invention, an onboard processing system for a low-orbit meteorological satellite comprises: a data exchange module, a preprocessing module, a resampling module and a product generation module;
[0071] The data exchange module receives and caches satellite multi-payload data and satellite platform data.
[0072] The data exchange module combines the cached payload data and platform data using a unified format framework. The unified format framework includes encapsulating the payload data into 144-byte packets, each of which contains a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data, and a 4-byte terminator.
[0073] The preprocessing module receives the combined data and separates, unpacks, and verifies it, positioning and calibrating the observation points. This separation, unpacking, and verification process includes: detecting the synchronization word on the received data stream, entering the synchronization state after the synchronization word is detected, identifying the payload based on the payload identifier following the synchronization word, and determining whether there has been packet loss based on packet counting. The payload data format is then checked for correctness, and the payload data is classified and stored if it passes the check.
[0074] The data exchange module receives the positioning and calibration results and forwards them to the resampling module.
[0075] The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time based on the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module. The generation of the standard spatiotemporal geographic grid includes: using satellite bus data as the time reference, assuming that the time series in the bus data output by the satellite is T1, T2, ..., and the first time received is T k , then T k The time starting point of the time-space reference grid is used as the satellite position corresponding to the first time as the spatial location. The spatial location is calculated every ΔT time interval after the first time; the time interval ΔT is the minimum value of the imaging cycle of each satellite payload. The resampling processing of the positioning and calibration results on the geographic grid includes: for each observation point, selecting 16 time-space reference points around the observation point for consideration, determining the point with the smallest distance, and associating the observation point with the point with the smallest distance; after the calculation is completed, checking the calibration values a1, a2, ..., a of each point associated with the time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
[0076]
[0077] The data exchange module receives the processing results of the resampling module and forwards them to the product generation module.
[0078] The product generation module performs product inversion on the processing results, generating various secondary products and forwarding the results to the data exchange module. This generation of the various secondary products utilizes a fully connected neural network, comprising an input layer, hidden layers, and an output layer, with 20 hidden layers. This neural network is trained using historical observations to obtain weights at each level. The input is the resampled values for each load channel, and the output is the predetermined inversion product.
[0079] The satellite platform sends a remote control command specifying the product type to be released, and the data exchange module selects the corresponding product data for output. The data exchange module can receive 64 product types and uses a 64-bit register to store product status. If each bit value is 1, the current product is released; if each bit value is 0, the current product is not released.
[0080] Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, and units provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same functions of the system and its various devices, modules, and units provided by the present invention in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, and units for implementing various functions can also be considered as both software modules implementing the method and structures within the hardware component.
[0081] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.
Claims
1. A low-orbit meteorological satellite onboard processing method, characterized in that: include: Step S1: The data exchange module receives and caches satellite multi-payload data and satellite platform data; Step S2: The data exchange module combines the cached payload data and platform data using a unified format framework; Step S3: The pre-processing module receives the combined data, separates, unpacks and verifies the data, and locates and calibrates the observation points therein; Step S4: The data exchange module receives the positioning and calibration results and forwards them to the resampling module; Step S5: The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time according to the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module; Step S6: The data exchange module receives the processing result of the resampling module and forwards it to the product generation module; Step S7: the product generation module performs product inversion on the processing results to generate various secondary products, and forwards the results to the data exchange module; Step S8: The satellite platform sends a remote control command to specify the type of product to be sent, and the data exchange module selects the corresponding product data for output.
2. The on-board processing method for low-orbit meteorological satellites according to claim 1, characterized in that: The step S2 includes encapsulating the payload data into packets of 144 bytes, each packet including a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data and a 4-byte terminator.
3. The on-board processing method for low-orbit meteorological satellites according to claim 1, characterized in that: The step S3 comprises: Step S3.1: Detect the synchronization word of the received data stream and enter the synchronization state after the synchronization word is detected; Step S3.2: Identify the payload according to the payload identifier after the synchronization word, and determine whether there is packet loss based on the packet count; Step S3.3: Check the correctness of the payload data format. If the correctness is found, the payload data is classified and stored.
4. The on-board processing method for low-orbit meteorological satellites according to claim 1, characterized in that: The generation of a standard spatiotemporal geographic grid includes: using satellite bus data as a time reference, assuming that the time series in the bus data output by the satellite is T1, T2, ..., and the first time received is T k , then T k The time starting point of the time-space reference grid is used as the satellite position corresponding to the first time as the spatial location, and the spatial location is calculated every ΔT time interval after the first time; the time interval ΔT is the minimum value of the imaging period of each satellite payload; The resampling process of the positioning and calibration results on the geographic grid includes: for each observation point, selecting 16 time-space reference points around the observation point for consideration, determining the point with the smallest distance, and associating the observation point with the point with the smallest distance; after the calculation is completed, checking the calibration values a1, a2, ..., a of the points associated with each time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
5. The on-board processing method for low-orbit meteorological satellites according to claim 1, characterized in that: The generation of various secondary products adopts a fully connected neural network, which includes an input layer, a hidden layer, and an output layer, wherein the number of hidden layers is 20; the fully connected neural network is trained using historical observations to obtain weights at each level; The input is the resampled value of each channel of each load, and the output is the predetermined inversion product; The step S8 includes: the data exchange module can receive 64 products, and uses a 64-bit register to store product status. If each bit value is 1, the current product is issued; If the value of each bit is 0, the current product will not be released.
6. A low-orbit meteorological satellite onboard processing system, characterized in that: include: Data exchange module, preprocessing module, resampling module and product generation module; The data exchange module receives and caches satellite multi-payload data and satellite platform data; The data exchange module combines the cached payload data and platform data using a unified format framework; The pre-processing module receives the combined data, separates, unpacks and verifies the data, and locates and calibrates the observation points. The data exchange module receives the positioning and calibration results and forwards them to the resampling module; The resampling module generates a standard spatiotemporal geographic grid covering a certain observation time based on the current satellite position and payload observation mode, and resamples the positioning and calibration results on the geographic grid, and forwards the processed results to the data exchange module; The data exchange module receives the processing results of the resampling module and forwards them to the product generation module; The product generation module performs product inversion on the processing results to generate various secondary products and forwards the results to the data exchange module; The satellite platform sends remote control instructions to specify the type of product to be issued, and the data exchange module selects the corresponding product data for output.
7. The onboard processing system for low-orbit meteorological satellites according to claim 6, characterized in that: The unified format framework includes encapsulating the payload data into packets of 144 bytes, each packet containing a 4-byte synchronization word, a 2-byte payload identifier, a 2-byte packet count, a 4-byte padding bit, 128 bytes of payload data, and a 4-byte terminator; It also includes: a dynamic backup module and a power supply module. The dynamic backup module realizes redundancy function and replaces the functions of the preprocessing module, resampling module and product generation module by uploading different software, and is used in the event of failure of the corresponding module; the power supply module realizes the conversion of the satellite's primary power supply into secondary power supply and supplies power to each module.
8. The low-orbit meteorological satellite onboard processing system according to claim 6, characterized in that: The separation, unpacking and verification processing includes: performing synchronization word detection on the received data stream, entering the synchronization state after detecting the synchronization word, then judging the payload according to the payload identifier after the synchronization word, and judging whether there is packet loss in combination with the packet count, and then checking the correctness of the payload data format. After the check is passed, the payload data is classified and stored.
9. The onboard processing system for low-orbit meteorological satellites according to claim 6, characterized in that: The generation of a standard spatiotemporal geographic grid includes: using satellite bus data as a time reference, assuming that the time series in the bus data output by the satellite is T1, T2, ..., and the first time received is T k , then T k The time starting point of the time-space reference grid is used as the satellite position corresponding to the first time as the spatial location, and the spatial location is calculated every ΔT time interval after the first time; the time interval ΔT is the minimum value of the imaging period of each satellite payload; The resampling process of the positioning and calibration results on the geographic grid includes: for each observation point, selecting 16 time-space reference points around the observation point for consideration, determining the point with the smallest distance, and associating the observation point with the point with the smallest distance; after the calculation is completed, checking the calibration values a1, a2, ..., a of the points associated with each time-space reference point. n and distance values r1, r2, …, r n (n≤10), remove the points with a distance of 0, and use the following formula to calculate the calibration value after resampling:
10. The onboard processing system for low-orbit meteorological satellites according to claim 6, characterized in that: The generation of various secondary products adopts a fully connected neural network, which includes an input layer, a hidden layer, and an output layer, wherein the number of hidden layers is 20; the fully connected neural network is trained using historical observations to obtain weights at each level; the input is the resampled value of each load and each channel, and the output is the predetermined inversion product; The data exchange module can receive 64 types of products and uses a 64-bit register to store product status. If each bit value is 1, the current product is issued; if each bit value is 0, the current product is not issued. The data exchange module uses anti-fuse FPGA and large-capacity FPGA.
Citation Information
Patent Citations
On-satellite real-time intelligent processing device for small general load data
CN115240053A
Cited By
Whole-process automatic meteorological satellite data processing method and related equipment
CN121412300A